Creation device, component control device, creation method, component control method and computer program

Patent metadata

Field	Value
Publication number	US-10967935-B2
Application number	US-201916453561-A
Country	US
Kind code	B2
Filing date	Jun 26, 2019
Priority date	Jul 9, 2018
Publication date	Apr 6, 2021
Grant date	Apr 6, 2021

Abstract

A creation device includes an acquisition part that acquires input information concerning traveling of a human-powered vehicle; and a creation part that creates different learning models that each produce output information concerning control of a component of the human-powered vehicle based on input information acquired by the acquisition part.

First claim

What is claimed is:

1. A creation device comprising:
at least one processor configured to execute operations including:
an acquisition part configured to acquire input information concerning traveling of a human-powered vehicle; and
a creation part configured to create different learning models that each produce output information concerning control of a component of the human-powered vehicle based on the input information acquired by the acquisition part.

2. The creation device according to claim 1 , wherein
the output information includes at least one of a gear stage and a gear ratio for a transmission.

3. The creation device according to claim 1 , wherein
the input information includes at least one of a traveling speed of the human-powered vehicle and a cadence of a crank of a driving mechanism.

4. The creation device according to claim 1 , wherein
the input information includes detection data of an attitude of the human-powered vehicle.

5. The creation device according to claim 1 , wherein
the input information includes detection data of a posture of a user riding the human-powered vehicle.

6. The creation device according to claim 1 , wherein
the input information includes detection data of a traveling environment of the human-powered vehicle.

7. The creation device according to claim 1 , wherein
the creation part is configured to input pieces of the input information in a different combination from a combination of information for another one of the learning models out of the pieces of the input information acquired by the acquisition part for each of the different learning models.

8. The creation device according to claim 1 , wherein
the creation part is configured to set pieces of the input information in a different combination for each of the different learning models.

9. The creation device according to claim 1 , wherein
the creation part is configured to assign weights different from those of another one of the learning models to a plurality of pieces of data included in the input information acquired by the acquisition part for each of the different learning models and inputs the plurality of pieces of the data that have been weighted.

10. The creation device according to claim 1 , wherein
the creation part is configured to set different weights for each of the different learning models.

11. The creation device according to claim 1 , wherein
the acquisition part is configured to acquire pieces of the input information at different points in time, and
the creation part is configured to input the pieces of the input information at the different points in time to each of the learning models.

12. The creation device according to claim 11 , wherein
an interval between the points in time when the pieces of the input information are acquired is different among the different learning models.

13. The creation device according to claim 1 , wherein
the acquisition part is configured to acquire pieces of the input information at different points in time, and the creation part is configured to input a variation between the pieces of the input information to the learning model.

14. The creation device according to claim 13 , wherein
a time duration corresponding to the variation between the pieces of the input information is different among the different learning models.

15. The creation device according to claim 1 , further comprising
a selection input that selects a targeted learning model out of the different learning models, wherein
the creation part creates the targeted learning model selected by the selection input.

16. The creation device according to claim 15 , wherein
the at least one processor further includes an evaluation part configured to evaluate the output information output from the targeted learning model selected by the selection input, and
the creation part is configured to update the targeted learning model based on an evaluation by the evaluation part.

17. The creation device according to claim 16 , further comprising
a user operable input device configured to accept a designation operation concerning the output information, and
the evaluation part being configured to make an evaluation by checking the output information output from the targeted learning model in response to an input of the input information acquired by the acquisition part to the targeted learning model against the designation operation accepted by the user operable input device.

18. The creation device according to claim 17 , wherein
the evaluation part is configured to provide the output information with a first evaluation concerning the targeted learning model upon determining the output information output from the targeted learning model and a detail of the designation operation do not match with each other.

19. The creation device according to claim 17 , wherein
the creation part is configured to update the targeted learning model by assigning a weight based on training data including a detail of the designation operation upon determining the output information output from the targeted learning model and the detail of the designation operation do not match with each other.

20. The creation device according to claim 17 , further comprising
a non-transitory computer readable storage that temporarily stores, one by one, a plurality of pieces of input information in time series for access by the acquisition part, and
the creation part being configured to set, upon acceptance of the designation operation by the user operable input device, the pieces of the input information acquired before and after a timing when the designation operation is performed as input data, and the creation part being configured to update the targeted learning model using the input data and a detail of the operation by the user operable input device.

21. The creation device according to claim 20 , wherein
the creation part is configured to update the targeted learning model by increasing a number of acquisitions of the input information upon acceptance of the designation operation by the user operable input device.

22. The creation device according to claim 20 , wherein
the creation part is configured to update the targeted learning model by increasing an acquisition frequency of input information upon acceptance of the designation operation by the user operable input device.

23. The creation device according to claim 16 , further comprising
an evaluation acceptance part configured to accept an evaluation provided by a user for action of the component based on the output information output from each of the different learning models, and
the targeted learning model being updated based on training data including a detail of the evaluation accepted by the evaluation acceptance part, the output information output from the targeted learning model at a timing upon acceptance of the evaluation, and the input information input in association with the output information.

24. The creation device according to claim 23 , further comprising
a user operable input device configured to accept a designation operation concerning the output information, and
the evaluation acceptance part being provided at the user operable input device or near the user operable input device so as to accept evaluation operation by the user.

25. The creation device according to claim 23 , wherein
the evaluation acceptance pa

Assignees

• Shimano Kk

Inventors

• Shimazu Hayato
• Takayama Hitoshi
• Shahana Satoshi
• Nishino Takafumi
• Nakajima Takehiko

Classifications

• B62J45/20 Primary

  Cycle computers as cycle accessories · CPC title

• B62M25/08 Primary

  with electrical or fluid transmitting systems · CPC title

• G06F18/285

  Selection of pattern recognition techniques, e.g. of classifiers in a multi-classifier system · CPC title

• G06F18/2155

  characterised by the incorporation of unlabelled data, e.g. multiple instance learning [MIL], semi-supervised techniques using expectation-maximisation [EM] or naïve labelling · CPC title

• G06F18/28

  Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries · CPC title